ASA PRESSROOM


Acoustical Society of America
159th Meeting Lay Language Papers


[ Lay Language Paper Index | Press Room ]



Enabling Hearing Impaired Listeners Better Communicate in Noisy Situations

 

Philipos C. Loizou - loizou@utdallas.edu

Department of Electrical Engineering

University of Texas-Dallas

Richardson, TX 75080

Popular version of paper 2aSC1

Presented Tuesday morning, April 20, 2010

159th ASA Meeting, Baltimore, MD.

Being able to focus on a conversation in a crowded room or noisy environment is a challenge for people who rely on hearing aids or cochlear implants. While these technological advances make it possible for many to hear who would not otherwise, they do not allow the individual to filter out background noise. In our work, we explored sophisticated algorithms to allow the microprocessors in these devices to filter noise in the same way that the unaided normal human ear does. Noise suppression algorithms in the past have been effective in improving speech quality, but have not improved speech intelligibility, particularly in noisy environments. In fact many of the noise suppression algorithms actually introduce speech distortion.

 

The approach taken in our study is based on the idea that because the characteristics of the noise differ across the various listening situations (e.g., train, car, restaurant, etc.) encountered by hearing-impaired listeners, the algorithms designed to suppress the noise ought to be optimized differently for each listening situation. Such an algorithm was developed in our lab and tested in three different types of environments which included multi-talker babble: restaurant noise, train noise and exhibition hall noise. Our study involved both normal hearing and hearing-impaired subjects wearing cochlear implant devices. Significant improvements in intelligibility, measured in terms of the number of words identified correctly, were obtained in our lab for both normal-hearing and hearing-impaired listeners. Our study demonstrated significant improvements in speech intelligibility when using environment-optimized algorithms, i.e., algorithms that are tuned to a particular listening environment. These preliminary results have been very promising, and the next step is to test these new algorithms in real-world environments using the patients own devices. These advanced algorithms and their application could mean a significant difference in the life of hearing impaired individuals who rely on hearing aids and cochlear implants to hear the world around them. It is our hope that this work could make it possible for hearing impaired listeners to better communicate in the noisy environments that are part of everyday life.